EURURO Vol. 72 No. 6

871 1030

(low [LDR] or high [HDR] dose rate), or radiotherapy
delivered with or without neoadjuvant/concurrent/
adjuvant ADT for the treatment of localised PCa
4. BT alone (LDR or HDR)
Studies reporting comparisons between any of the above
treatments were included. Studies reporting within-treat-
ment comparisons only (eg, radiotherapy vs radiotherapy in
combination with ADT) were excluded, as the aim of this
review was to compare QoL outcomes between different
management options.
2.5. Types of outcome measures included
The primary outcome of this review was specified a
priori and included cancer-specific QoL after primary
cancer treatment, assessed by a validated PROM as
defined by authors. In addition, a list of validated PROMs
was used to identify potentially relevant studies and
incorporated into the search strategy . Studies using
non–cancer-specific PROMs or studies reporting cancer-
specific QoL using a nonvalidated tool were excluded, as
were those studies reporting incomplete data from
subdomains of QoL PROMs. Secondary outcomes were
subdomains of QoL PROMs related to PCa QoL, for
example, sexual function, urinary function, and bowel
function.
2.6. Assessment of risk of bias and confounding
Randomised controlled trial (RCT) risk of bias (RoB)
assessment was undertaken using the recommended tool
in the Cochrane Handbook for Systematic Reviews of
Interventions
In nonrandomised comparative studies (NRCSs), RoB was
assessed using additional domains to assess the risk of
confounders, which were developed a priori with clinical
content experts (EAU Prostate Cancer Guideline Panel). This
is a pragmatic approach informed by methodological
literature pertaining to assessing RoB in NRCSs . The
main confounding factors identified included baseline QoL
score, age, comorbidities (any classification), and baseline
Gleason score.
2.7. Data analysis
A data extraction form was developed a priori to collect
information on study design, participant demographics,
characteristics of interventions, and outcome measures.
Two reviewers independently extracted data relating to the
prespecified outcomes. Descriptive statistics were used to
summarise baseline characteristic data. For studies with
more than two intervention groups, only the intervention
groups relevant to the review were selected. If relevant
data could be extracted and it was appropriate to do so, a
meta-analysis of RCT data was planned. For studies with
multiple publications, only the most up-to-date or complete
data for each outcome were utilised. If meta-analyses of
RCTs were inappropriate, a narrative synthesis of the
evidence was performed. A narrative synthesis was
undertaken for NRCSs .
3. Evidence synthesis
3.1. Quantity of evidence identified
The study selection process is outlined in A total of
18 studies were eligible for inclusion: three RCTs
and 15 NRCSs of which one NRCS had
multiple publications. Ultimately, a total of 13 604 patients
were recruited (2011 from RCTs and 11 593 from NRCSs).
3.2. Characteristics of the included studies
present the baseline study characteristics for
the three RCTs and 15 NRCSs, respectively. Owing to
heterogeneity of study PROM data, a meta-analysis was not
performed, and consequently, data were summarised
narratively instead.
3.3. RoB and quality assessment of the included studies
A and 2B present the RoB summary and
confounder assessment for the three RCTs and
15 NRCSs . As it was not possible to blind the
participants to their intervention, all RCTs had a
high RoB for blinding of participants, but we did not judge
that this necessarily compromised study quality. One RCT
was also judged to have a high selection bias (as only
19% of patients were randomly assigned to treatment arms),
high bias due to closing prematurely, and unclear detection
bias, while another was judged to have unclear
selection and funding biases.
The NRCSs had a high risk of selection, performance, and
detection biases. The risks of reporting bias were low, while
those of attrition bias were moderate. All confounders were
measured and corrected for, in five studies
3.4. Comparisons of interventions results
3.4.1. Data from RCTs
Statistically significant differences for QoL outcomes
between or within treatment groups at the latest follow-
up of each RCT are shown in The complete
summary of the outcome results can be found in
Supplementary Table 1.
3.4.1.1. RP versus EBRT versus active monitoring. Data were
obtained from the recently published Prostate Testing for
Cancer and Treatment (ProtecT) trial , where 1643 men
were randomised to active monitoring, RP, or EBRT. The trial
predominantly enrolled men with low- and intermediate-
risk PCa. Trial retention and completion of follow-up
assessments were >85% for most outcome measures.
Analyses were performed according to an intent-to-treat
basis. Approximately 50% (291) of men who initially
underwent active monitoring had either surgery or
radiotherapy by the end of November 2015.
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 8 6 9 – 8 8 5 871